something to munch on (Was: Re: [Stackless] cool stuff)

Sam M. Rushing srushing at ironport.com
Thu Aug 15 23:13:32 CEST 2002


> -----Original Message-----
> From: Paul Sheer [mailto:psheer at icon.co.za]
> Sent: Thursday, August 15, 2002 4:04 AM
> To: Ctismer (forward); stackless at tismer.com; marc at welz.org.za
> Subject: something to munch on (Was: Re: [Stackless] cool stuff)

> The point here is that we are moving from an event
> based model toward a thread model. It used to be an
> "either thread based or event based". But now we are
> somewhere in between. So my question is: at what point
> do we cross the line and end up with the same problems
> that traditional multithreading has?

At IronPort we've been living with this model for about
eighteen months.  Once your system reaches a certain level
of complexity you start seeing traditional 'thread race'
conditions.  We've never run up against 'mystery deadlocks',
because we have such clear visibility into what each thread
is doing (we can get a traceback on any and all threads), but
every once in a while there'll be a 'hey, we're multi-threaded!'
surprise.

The best way to control this is to be aware of exactly
which parts of your system might trigger a context
switch (typically, just network I/O).  I think the best
approach for us has been 'strategic serialization'.

-Sam
_______________________________________________
Stackless mailing list
Stackless at www.tismer.com
http://www.tismer.com/mailman/listinfo/stackless



More information about the Stackless mailing list